In the age of e-Commerce, customers increasingly rely on product reviews to make informed decisions. However, the vast number of reviews on different e-commerce platforms such as Amazon and Flip-kart, etc. can be overwhelming. We propose an efficient Product Review Summarization technique using a Retrieval Augmented Generation (RAG) approach to address this. RAG retrieves relevant information from a vector store and uses a Large Language Model (LLM) to generate coherent summaries. Our method enhances traditional RAG systems by implementing ensemble retrievers that combine multiple retrieval models to ensure reliability. Additionally, we explore a hybrid RAG architecture with data compression and cross-encoder re-ranking to improve summary faithfulness. We evaluated our approach using metrics such as SummaC, CTC, FactCC, and FactGraph, demonstrating improved faithfulness scores over standard retrieval methods.